2edvol1c

7/28/2019 2edvol1c

1/65

3. Chemical aspects

3.1 Background information used

The assessment of the toxicity of drinking-water contaminants has been made on the basis ofpublished reports from the open literature, information submitted by governments and otherinterested parties, and unpublished proprietary data. In the development of the guideline values,existing international approaches to developing guidelines were carefully considered. Previousrisk assessments developed by the International Programme on Chemical Safety (IPCS) inEnvironmental Health Criteria monographs, the International Agency for Research on Cancer(IARC), the Joint FAO/WHO Meetings on Pesticide Residues (JMPR), and the Joint FAO/WHOExpert Committee on Food Additives (JECFA) were reviewed. These assessments were reliedupon except where new information justified a reassessment. The quality of new data wascritically evaluated prior to their use in risk assessment.

3.2 Drinking-water consumption and body weight

Global data on the consumption of drinking-water are limited. In studies carried out in Canada,

the Netherlands, the United Kingdom, and the United States of America, the average daily percapita consumption was usually found to be less than 2 litres, but there was considerablevariation between individuals. As water intake is likely to vary with climate, physical activity, andculture, the above studies, which were conducted in temperate zones, can give only a limitedview of consumption patterns throughout the world. At temperatures above 25 C, for example,there is a sharp rise in fluid intake, largely to meet the demands of an increased sweat rate.

In developing the guideline values for potentially hazardous chemicals, a daily per capitaconsumption of 2 litres by a person weighing 60 kg was generally assumed. The guideline valuesset for drinking-water using this assumption do, on average, err on the side of caution. However,such an assumption may underestimate the consumption of water per unit weight, and thusexposure, for those living in hot climates as well as for infants and children, who consume morefluid per unit weight than adults.

The higher intakes, and hence exposure, for infants and children apply for only a limited time, butthis period may coincide with greater sensitivity to some toxic agents and less for others.Irreversible effects that occur at a young age will have more social and public health significancethan those that are delayed. Where it was judged that this segment of the population was at aparticularly high risk from exposure to certain chemicals, the guideline value was derived on thebasis of a 10-kg child consuming 1 litre per day or a 5-kg infant consuming 0.75 litre per day. Thecorresponding daily fluid intakes are higher than for adults on a body weight basis.

3.3 Inhalation and dermal absorption

The contribution of drinking-water to daily exposure includes direct ingestion as well as some

indirect routes, such as inhalation of volatile substances and dermal contact during bathing orshowering.

In most cases, the data were insufficient to permit reliable estimates of exposure by inhalationand dermal absorption of contaminants present in drinking-water. It was not possible, therefore,to address intake from these routes specifically in the derivation of the guideline values. However,that portion of the total tolerable daily intake (TDI) allocated to drinking-water is generallysufficient to allow for these additional routes of intake (see section 3.4.1). When there is concernthat potential inhalation of volatile compounds and dermal exposure from various indoor wateruses (such as showering) are not adequately addressed, authorities could adjust the guideline

7/28/2019 2edvol1c

2/65

value.

3.4 Health risk assessment

There are two principal sources of information on health effects resulting from exposure tochemicals that can be used in deriving guideline values. The first is studies on humanpopulations. The value of such investigations is often limited, owing to lack of quantitativeinformation on the concentrations to which people are exposed or on simultaneous exposure toother agents. The second, and the one used most often, is toxicity studies using laboratoryanimals. Such studies are generally limited because of the relatively small numbers of animalsused and the relatively high doses administered. Furthermore, there is a need to extrapolate theresults to the low doses to which human populations are usually exposed.

In order to derive a guideline value to protect human health, it is necessary to select the mostsuitable experimental animal study on which to base the extrapolation. Data from well-conductedstudies, where a clear dose - response relationship has been demonstrated, are preferred. Expertjudgement was exercised in the selection of the most appropriate study from the range ofinformation available.

3.4.1 Derivation of guideline values using a tolerable daily intake approach

For most kinds of toxicity, it is generally believed that there is a dose below which no adverseeffects will occur. For chemicals that give rise to such toxic effects, a tolerable daily intake (TDI)can be derived as follows:

UF

LOAELorNOAELTDI =

where:

NOAEL = no-observed-adverse-effect level,LOAEL = lowest-observed-adverse-effect level,UF= uncertainty factor.

The guideline value (GV) is then derived from the TDI as follows:

C

PbwTDIGV

=

where:

bw= body weight (60 kg for adults, 10 kg for children, 5 kg for infants),P= fraction of the TDI allocated to drinking-water,C = daily drinking-water consumption (2 litres for adults, 1 litre for children, 0.75 litre for

infants).

Tolerable daily intake

The TDI is an estimate of the amount of a substance in food or drinking-water, expressed on abody weight basis (mg/kg or g/kg of body weight), that can be ingested daily over a lifetimewithout appreciable health risk.

Over many years, JECFA and JMPR have developed certain principles in the derivation ofacceptable daily intakes (ADIs). These principles have been adopted where appropriate in the

7/28/2019 2edvol1c

3/65

derivation of TDIs used in developing guideline values for drinking-water quality.

ADIs are established for food additives and pesticide residues that occur in food for necessarytechnological purposes or plant protection reasons. For chemical contaminants, which usuallyhave no intended function in drinking-water, the term tolerable daily intake is seen as moreappropriate than acceptable daily intake, as it signifies permissibility rather than acceptability.

As TDIs are regarded as representing a tolerable intake for a lifetime, they are not so precise thatthey cannot be exceeded for short periods of time. Short-term exposure to levels exceeding theTDI is not a cause for concern, provided the individuals intake averaged over longer periods oftime does not appreciably exceed the level set. The large uncertainty factors generally involved inestablishing a TDI (see below) serve to provide assurance that exposure exceeding the TDI forshort periods is unlikely to have any deleterious effects upon health. However, considerationshould be given to any potential acute toxic effects that may occur if the TDI is substantiallyexceeded for short periods of time.

The calculated TDI was used to derive the guideline value, which was then rounded to onesignificant figure. In some instances, ADI values with only one significant figure set by JECFA orJMPR were used to calculated the guideline value. The guideline value was generally rounded toone significant figure to reflect the uncertainty in animal toxicity data and exposure assumptions

made. More than one significant figure was used for guideline values only where extensiveinformation on toxicity and exposure to humans provided greater certainty.

No-observed-adverse-effect level and lowest-observed-adverse-effect level

The NOAEL is defined as the highest dose or concentration of a chemical in a single study, foundby experiment or observation, that causes no detectable adverse health effect. Wheneverpossible, the NOAEL is based on long-term studies, preferably of ingestion in drinking-water.However, NOAELs obtained from short-term studies and studies using other sources of exposure(e.g., food, air) may also be used.

If a NOAEL is not available, a LOAEL may be used, which is the lowest observed dose orconcentration of a substance at which there is a detectable adverse health effect. When a LOAEL

is used instead of a NOAEL, an additional uncertainty factor is normally used (see below).

Uncertainty factors

The application of uncertainty factors has been widely used in the derivation of ADIs for foodadditives, pesticides, and environmental contaminants. The derivation of these factors requiresexpert judgement and a careful sifting of the available scientific evidence.

In the derivation of the WHO drinking-water quality guideline values, uncertainty factors wereapplied to the lowest NOAEL or LOAEL for the response considered to be the most biologicallysignificant and were determined by consensus among a group of experts using the approachoutlined below:

Source of uncertainty Factor Interspecies variation (animals to humans) 1-10Intraspecies variation (individual variations) 1-10Adequacy of studies or database 1-10Nature and severity of effect 1-10

Inadequate studies or databases include those that used a LOAEL instead of a NOAEL andstudies considered to be shorter in duration than desirable. Situations in which the nature orseverity of effect might warrant an additional uncertainty factor include studies in which the end-point was malformation of a fetus or in which the end-point determining the NOAEL was directly

7/28/2019 2edvol1c

4/65

related to possible carcinogenicity. In the latter case, an additional uncertainty factor was appliedfor carcinogenic compounds for which a guideline value was derived using a TDI approach (seesection 3.4.2). Factors lower than 10 were used, for example, for interspecies variation whenhumans are known to be less sensitive than the animal species studied.

The total uncertainty factor should not exceed 10 000. If the risk assessment would lead to ahigher uncertainty factor, then the resulting TDI would be so imprecise as to lack meaning. Forsubstances for which uncertainty factors were greater than 1000, guideline values are designatedas provisional in order to emphasize the high level of uncertainty inherent in these values.

The selection and application of uncertainty factors are important in the derivation of guidelinevalues for chemicals, as they can make a considerable difference to the values set. Forcontaminants for which there is relatively little uncertainty, the guideline value was derived usinga small uncertainty factor. For most contaminants, however, there is great scientific uncertainty,and a large uncertainty factor was used. Hence, there may be a large margin of safety above theguideline value before adverse health effects result.

There is considerable merit in using a method that allows a high degree of flexibility. However, itis important that, where possible, the derivation of the uncertainty factor used in calculating aguideline value is clearly presented as part of the rationale. This helps authorities in using the

guidelines, as the safety margin in allowing for local circumstances is clear. It also helps indetermining the urgency and nature of the action required in the event that a guideline value isexceeded.

Allocation of intake

Drinking-water is not usually the sole source of human exposure to the substances for whichguideline values have been set. In many cases, the intake from drinking-water is small incomparison with that from other sources such as food and air. Guideline values derived using theTDI approach take into account exposure from all sources by apportioning a percentage of theTDI to drinking-water. This approach ensures that total daily intake from all sources (includingdrinking-water containing concentrations of the substance at or near the guideline value) does notexceed the TDI.

Wherever possible, data concerning the proportion of total intake normally ingested in drinking-water (based on mean levels in food, air, and drinking-water) or intakes estimated on the basis ofconsideration of physical and chemical properties were used in the derivation of the guidelinevalues. Where such information was not available, an arbitrary (default) value of 10% for drinking-water was used. This default value is, in most cases, sufficient to account for additional routes ofintake (i.e., inhalation and dermal absorption) of contaminants in water.

It is recognized that exposure from various media may vary with local circumstances. It should beemphasized, therefore, that the derived guideline values apply to a typical exposure scenario orare based on default values that may not be applicable for all areas. In those areas whererelevant data on exposure are available, authorities are encouraged to develop context-specificguideline values that are tailored to local circumstances and conditions. For example, in areas

where the intake of a particular contaminant in drinking-water is known to be much greater thanthat from other sources (i.e., air and food), it may be appropriate to allocate a greater proportionof the TDI to drinking-water to derive a guideline value more suited to the local conditions. Inaddition, in cases in which guideline values are exceeded, efforts should be made to assess thecontribution of other sources to total intake; if practicable, exposure from these sources should beminimized.

3.4.2 Derivation of guideline values for potential carcinogens

The evaluation of the potential carcinogenicity of chemical substances is usually based on long-

7/28/2019 2edvol1c

5/65

term animal studies. Sometimes data are available on carcinogenicity in humans, mostly fromoccupational exposure.

On the basis of the available evidence, IARC categorizes chemical substances with respect totheir potential carcinogenic risk into the following groups (for a detailed description of theclassifications, see box below):

Evaluation of carcinogenic risk to humans

IARC considers the body of evidence as a whole in order to reach an overall evaluation of thecarcinogenicity for humans of an agent, mixture, or circumstance of exposure.

The agent, mixture, or exposure circumstance is described according to the wording of one of thefollowing categories, and the designated group is given. The categorization of an agent, mixture,or exposure circumstance is a matter of scientific judgement, reflecting the strength of theevidence derived from studies in humans and in experimental animals and from other relevantdata.

Group 1. The agent (mixture) is carcinogenic to humans.The exposure circumstance entails exposures that are carcinogenic to humans.

This category is used when there is sufficient evidence of carcinogenicity in humans.Exceptionally, an agent (mixture) may be placed in this category when evidence in humans isless than sufficient but there is sufficient evidence of carcinogenicity in experimental animals andstrong evidence in exposed humans that the agent (mixture) acts through a relevant mechanismof carcinogenicity.

Group 2

This category includes agents, mixtures, and exposure circumstances for which, at one extreme,the degree of evidence of carcinogenicity in humans is almost sufficient, as well as those forwhich, at the other extreme, there are no human data but for which there is evidence ofcarcinogenicity in experimental animals. Agents, mixtures, and exposure circumstances are

assigned to either group 2A (probably carcinogenic to humans) or group 2B (possiblycarcinogenic to humans) on the basis of epidemiological and experimental evidence ofcarcinogenicity and other relevant data.

Group 2A. The agent (mixture) is probably carcinogenic to humans.The exposure circumstance entails exposures that are probably carcinogenic to humans.

This category is used when there is limited evidence of carcinogenicity in humans and sufficientevidence of carcinogenicity in experimental animals. In some cases, an agent (mixture) may beclassified in this category when there is inadequate evidence of carcinogenicity in humans andsufficient evidence of carcinogenicity in experimental animals and strong evidence that thecarcinogenesis is mediated by a mechanism that also operates in humans. Exceptionally, anagent, mixture, or exposure circumstance may be classified in this category solely on the basis of

limited evidence of carcinogenicity in humans.

Group 2B. The agent (mixture) is possibly carcinogenic to humans.The exposure circumstance entails exposures that are possibly carcinogenic to humans.

This category is used for agents, mixtures, and exposure circumstances for which there is limitedevidence of carcinogenicity in humans and less than sufficient evidence of carcinogenicity inexperimental animals. It may also be used when there is inadequate evidence of carcinogenicityin humans but there is sufficient evidence of carcinogenicity in experimental animals. In someinstances, an agent, mixture, or exposure circumstance for which there is inadequate evidence of

7/28/2019 2edvol1c

6/65

carcinogenicity in humans but limited evidence of carcinogenicity in experimental animalstogether with supporting evidence from other relevant data may be placed in this group.

Group 3. The agent (mixture or exposure circumstance) is not classifiable as to itscarcinogenicity to humans.

This category is used most commonly for agents, mixtures, and exposure circumstances forwhich the evidence of carcinogenicity is inadequate in humans and inadequate or limited inexperimental animals.

Exceptionally, agents (mixtures) for which the evidence of carcinogenicity in inadequate inhumans but sufficient in experimental animals may be placed in this category when there isstrong evidence that the mechanism of carcinogenicity in experimental animals does not operatein humans.

Agents, mixtures, and exposure circumstances that do not fall into any other group are alsoplaced in this category.

Group 4. The agent (mixture) is probably not carcinogenic to humans.

This category is used for agents or mixtures for which there is evidence suggesting lack ofcarcinogenicityin humans and in experimental animals. In some instances, agents or mixtures forwhich there is inadequate evidence of carcinogenicity in humans but evidence suggesting lack ofcarcinogenicityin experimental animals, consistently and strongly supported by a broad range ofother relevant data, may be classified in this group.

Group 1: the agent is carcinogenic to humansGroup 2A: the agent is probably carcinogenic to humansGroup 2B: the agent is possibly carcinogenic to humansGroup 3: the agent is not classifiable as to its carcinogenicity to humansGroup 4: the agent is probably not carcinogenic to humans.

In establishing the present guideline values for drinking-water quality, the IARC classification for

carcinogenic compounds was taken into consideration. For a number of compounds, additionalinformation was also available.

It is generally considered that the initiating event in the process of chemical carcinogenesis is theinduction of a mutation in the genetic material (DNA) of somatic cells (i.e., cells other than ova orsperm). Because the genotoxic mechanism theoretically does not have a threshold, there is aprobability of harm at any level of exposure. Therefore, the development of a TDI is consideredinappropriate, and mathematical low-dose extrapolation is applied. On the other hand, there arecarcinogens that are capable of producing tumours in animals or humans without exerting agenotoxic activity, but acting through an indirect mechanism. It is generally believed that athreshold dose exists for these non-genotoxic carcinogens.

In order to make the distinction with respect to the underlying mechanism of carcinogenicity, each

compound that has been shown to be a carcinogen was evaluated on a case-by-case basis,taking into account the evidence of genotoxicity, the range of species affected, and the relevanceto humans of the tumours observed in experimental animals.

For carcinogens for which there is convincing evidence to suggest a non-genotoxic mechanism,guideline values were calculated using a TDI approach, as described in section 3.4.1.

In the case of compounds considered to be genotoxic carcinogens, guideline values weredetermined using a mathematical model, and the guideline values are the as concentration indrinking-water associated with an estimated upper bound excess lifetime cancer risk of 10

-5(one

7/28/2019 2edvol1c

7/65

additional cancer per 100 000 of the population ingesting drinking-water containing the substanceat the guideline value for 70 years). Concentrations associated with estimated excess lifetimecancer risks of 10

-4and 10

-6can be calculated by multiplying and dividing, respectively, the

guideline value by 10. These values are also presented in this volume to emphasize the fact thateach country should select its own appropriate risk level. In cases in which the concentrationassociated with a 10

-5excess lifetime cancer risk is not practical because of inadequate analytical

or treatment technology, a provisional guideline value was set at a practicable level and theestimated associated cancer risk presented.

Although several models exist, the linearized multistage model was generally adopted in thedevelopment of these guidelines. As indicated in Volume 2, other models were considered moreappropriate in a few cases.

It should be emphasized, however, that guideline values for carcinogenic compounds computedusing mathematical models must be considered at best as a rough estimate of the cancer risk.These models do not usually take into account a number of biologically important considerations,such as pharmacokinetics, DNA repair, or immunological protection mechanisms. However, themodels used are conservative and probably err on the side of caution.

To account for differences in metabolic rates between experimental animals and humans - the

former are more closely correlated with the ratio of body surface areas than with body weights - asurface area to body weight correction is sometimes applied to quantitative estimates of cancerrisk derived on the basis of models for low-dose extrapolation. Incorporation of this factorincreases the risk by approximately one order of magnitude (depending on the species uponwhich the estimate is based) and increases the risk estimated on the basis of studies in micerelative to that in rats. The incorporation of this factor is considered to be overly conservative,particularly in view of the fact that linear extrapolation most likely overestimates risk at low doses;indeed, Crump et al. concluded that all measures of dose except dose rate per unit of bodyweight tend to result in overestimation of human risk.

1Consequently, guideline values for

carcinogenic contaminants were developed on the basis of quantitative estimates of risk thatwere not corrected for the ratio of surface area to body weight.

1Crump K, Allen B, Shipp A. Choice of dose measures for extrapolating carcinogenic risk from

animals to humans: an empirical investigation of 23 chemicals. Health physics, 1989, 57,Suppl. 1: 387-393

3.5 Mixtures

Chemical contaminants of drinking-water supplies are present with numerous other inorganic andorganic constituents. The guideline values were calculated separately for individual substances,without specific consideration of the potential for interaction of each substance with othercompounds present. However, the large margin of safety incorporated in the majority of guidelinevalues is considered to be sufficient to account for potential interactions. In addition, the majorityof contaminants will not be present at concentrations at or near their guideline value.

There may, however, be occasions when a number of contaminants with similar toxicologicaleffects are present at levels near their respective guideline values. In such cases, decisionsconcerning appropriate action should be made, taking into consideration local circumstances.Unless there is evidence to the contrary, it is appropriate to assume that the toxic effects of thesecompounds are additive.

3.6 Summary statements

3.6.1 Inorganic constituents

7/28/2019 2edvol1c

8/65

Aluminium

Aluminium is a widespread and abundant element, comprising some 8% of the earth's crust.Aluminium compounds are widely used as coagulants in treatment of water for public supply andthe presence of aluminium in drinking-water is frequently due to deficiencies in the control andoperation of the process. Human exposure may occur by a variety of routes, with drinking-waterprobably contributing less than 5% of the total intake.

The metabolism of aluminium in humans is not well understood, but it appears that inorganicaluminium is poorly absorbed and that most of the absorbed aluminium is rapidly excreted in theurine.

Aluminium is of low toxicity in laboratory animals, and JECFA established a provisional tolerableweekly intake (PTWI) of 7 mg/kg of body weight in 1988. However, this was based on studies ofaluminium phosphate (acidic); the chemical form of aluminium in drinking-water is different.

In some studies, aluminium has appeared to be associated with the brain lesions characteristic ofAlzheimer disease, and in several ecological epidemiological studies the incidence of Alzheimerdisease has been associated with aluminium in drinking-water. These ecological analyses must

be interpreted with caution and should be confirmed in analytical epidemiological studies.

There is a need for further studies, but the balance of epidemiological and physiological evidenceat present does not support a causal role for aluminium in Alzheimer disease. Therefore, nohealth-based guideline value is recommended. However, a concentration of aluminium of 0.2mg/litre in drinking-water provides a compromise between the practical use of aluminium salts inwater treatment and discoloration of distributed water.

Ammonia

The term ammonia includes the non-ionized (NH3) and ionized (NH4+) species. Ammonia in the

environment originates from metabolic, agricultural, and industrial processes and fromdisinfection with chloramine. Natural levels in ground and surface waters are usually below 0.2

mg/litre. Anaerobic ground waters may contain up to 3 mg/litre. Intensive rearing of farm animalscan give rise to much higher levels in surface water. Ammonia contamination can also arise fromcement mortar pipe linings. Ammonia in water is an indicator of possible bacterial, sewage, andanimal waste pollution.

Ammonia is a major component of the metabolism of mammals. Exposure from environmentalsources is insignificant in comparison with endogenous synthesis of ammonia. Toxicologicaleffects are observed only at exposures above about 200 mg/kg of body weight.

Ammonia in drinking-water is not of immediate health relevance, and therefore no health-basedguideline value is proposed. However, ammonia can compromise disinfection efficiency, result innitrite formation in distribution systems, cause the failure of filters for the removal of manganese,and cause taste and odour problems.

Antimony

Antimony salts and possibly organic complexes of antimony are typically found in food and waterat low levels. Reported concentrations of antimony in drinking-water are usually less than 4g/litre. Estimated dietary intake for adults is about 0.02 mg/day. Where antimony-tin solder isbeginning to replace lead solder, exposure to antimony may increase in the future.

In its overall evaluation based on inhalation exposure, IARC concluded that antimony trioxide ispossibly carcinogenic to humans (Group 2B) and antimony trisulfide is not classifiable as to its

7/28/2019 2edvol1c

9/65

carcinogenicity in humans (Group 3).

In a limited lifetime study in which rats were exposed to antimony in drinking-water at a singledose level of 0.43 mg/kg of body weight per day, effects observed were decreased longevity andaltered blood levels of glucose and cholesterol. No effects were observed on the incidence ofbenign or malignant tumours.

An uncertainty factor of 500 (100 for inter- and intraspecies variation and 5 for the use of aLOAEL instead of a NOAEL) was applied to the LOAEL of 0.43 mg/kg of body weight per day,giving a TDI of 0.86 g/kg of body weight. An allocation of 10% of the TDI to drinking-water givesa concentration of 0.003 mg/litre (rounded figure), which is below the limit of practical quantitativeanalysis. The provisional guideline value for antimony has therefore been set at a practicalquantification level of 0.005 mg/litre. This results in a margin of safety of approximately 250-foldfor potential health effects, based on the LOAEL of 0.43 mg/kg of body weight per day observedin the limited lifetime study in rats.

Arsenic

Arsenic is widely distributed throughout the earth's crust and is used commercially, primarily inalloying agents. It is introduced into water through the dissolution of minerals and ores, from

industrial effluents, and from atmospheric deposition; concentrations in ground water in someareas are sometimes elevated as a result of erosion from natural sources. The average dailyintake of inorganic arsenic in water is estimated to be similar to that from food; intake from air isnegligible.

Inorganic arsenic is a documented human carcinogen and has been classified by IARC in Group1. A relatively high incidence of skin and possibly other cancers that increase with dose and agehas been observed in populations ingesting water containing high concentrations of arsenic.

Arsenic has not been shown to be carcinogenic in the limited bioassays in animal species that areavailable, but it has given positive results in studies designed to assess the potential for tumourpromotion. Arsenic has not been shown to be mutagenic in bacterial and mammalian assays,although it has been shown to induce chromosomal aberrations in a variety of cultured cell types,

including human cells; such effects have not been observed in vivo.

Data on the association between internal cancers and ingestion of arsenic in drinking-water wereinsufficient for quantitative assessment of risk. Instead, owing to the documented carcinogenicityof arsenic in drinking-water in human populations, the lifetime risk of skin cancer was estimatedusing a multistage model. On the basis of observations in a population ingesting arsenic-contaminated drinking-water, the concentration associated with an excess lifetime skin cancerrisk of 10

-5was calculated to be 0.17 g/litre. This value may, however, overestimate the actual

risk of skin cancer owing to the possible contribution of other factors to disease incidence in thepopulation and to possible dose-dependent variations in metabolism that could not be taken intoconsideration. In addition, this value is below the practical quantification limit of 10 g/litre.

With a view to reducing the concentration of this carcinogenic contaminant in drinking-water, a

provisional guideline value for arsenic in drinking-water of 0.01 mg/litre is established. Theestimated excess lifetime skin cancer risk associated with exposure to this concentration is 6 10

-4.

A similar value may be derived (assuming a 20% allocation to drinking-water) on the basis of theprovisional maximum tolerable daily intake (PMTDI) for inorganic arsenic of 2 g/kg of bodyweight established by JECFA in 1983 and confirmed as a PTWI of 15 g/kg of body weight forinorganic arsenic in 1988. JECFA noted, however, that the margin between the PTWI and intakesreported to have toxic effects in epidemiological studies was narrow.

7/28/2019 2edvol1c

10/65

Asbestos

Asbestos is introduced into water by the dissolution of asbestos-containing minerals and ores aswell as from industrial effluents, atmospheric pollution, and asbestos-cement pipes in thedistribution system. Exfoliation of asbestos fibres from asbestos-cement pipes is related to theaggressiveness of the water supply. Limited data indicate that exposure to airborne asbestosreleased from tapwater during showers or humidification is negligible.

Asbestos is a known human carcinogen by the inhalation route. Although well studied, there hasbeen little convincing evidence of the carcinogenicity of ingested asbestos in epidemiologicalstudies of populations with drinking-water supplies containing high concentrations of asbestos.Moreover, in extensive studies in animal species, asbestos has not consistently increased theincidence of tumours of the gastrointestinal tract. There is, therefore, no consistent evidence thatingested asbestos is hazardous to health, and thus it was concluded that there was no need toestablish a health-based guideline value for asbestos in drinking-water.

Barium

Barium occurs as a number of compounds in the earth's crust and is used in a wide variety ofindustrial applications, but it is present in water primarily from natural sources. In general, food is

the principal source of exposure to barium; however, in areas where barium concentrations inwater are high, drinking-water may contribute significantly to total intake. Intake from air isnegligible.

Although an association between mortality from cardiovascular disease and the barium content ofdrinking-water was reported in an ecological epidemiological study, these results were notconfirmed in an analytical epidemiological study of the same population. Moreover, in a short-term study in a small number of volunteers, there was no consistent indication of adversecardiovascular effects following exposure to barium at concentrations of up to 10 mg/litre in water.There was, however, an increase in the systolic blood pressure of rats exposed to relatively lowconcentrations of barium in drinking-water.

A guideline value of 0.7 mg/litre (rounded figure) was derived using the NOAEL of 7.3 mg/litre

from the most sensitive epidemiological study conducted to date, in which there were nosignificant differences in blood pressure or the prevalence of cardiovascular disease between apopulation drinking water containing a mean barium concentration of 7.3 mg/litre and oneingesting water containing barium at 0.1 mg/litre, and incorporating an uncertainty factor of 10 toaccount for intraspecies variation.

This value is close to that derived on the basis of the results of toxicological studies in animalspecies. A TDI of 51 g/kg of body weight was calculated, based on a NOAEL of 0.51 mg/kg ofbody weight per day in a chronic study in rats and incorporating uncertainty factors of 10 forintraspecies variation and 1 for interspecies variation, as the results of a well-conductedepidemiological study indicate that humans are not more sensitive than rats to barium in drinking-water. The value derived from this TDI based on 20% allocation to drinking-water would be 0.3mg/litre (rounded figure).

The guideline value for barium in drinking-water is 0.7 mg/litre.

Beryllium

Beryllium has a number of important minor uses, based mostly on its heat resistance. It is foundinfrequently in drinking-water and only at very low concentrations, usually less than 1 g/litre.

Beryllium appears to be poorly absorbed from the gastrointestinal tract. Beryllium and berylliumcompounds have been classified by IARC as being probably carcinogenic to humans (Group 2A)

7/28/2019 2edvol1c

11/65

on the basis of occupational exposure and inhalation studies in laboratory animals. There are noadequate studies by which to judge whether it is carcinogenic by oral exposure.

Beryllium has been shown to interact with DNA and cause gene mutations, chromosomalaberrations, and sister chromatid exchange in cultured mammalian somatic cells, although it hasnot been shown to be mutagenic in bacterial test systems.

There are no suitable oral data on which to base a toxicologically supportable guideline value.However, the very low concentrations of beryllium normally found in drinking-water seem unlikelyto pose a hazard to consumers.

Boron

Elemental boron is used principally in composite structural materials, and boron compounds areused in some detergents and industrial processes. Boron compounds are released into waterfrom industrial and domestic effluents. Boron is usually present in drinking-water atconcentrations of below 1 mg/litre, but some higher levels have been found as a result ofnaturally occurring boron. The total daily intake of boron is estimated to be between 1 and 5 mg.

When administered as borate or boric acid, boron is rapidly and almost completely absorbed from

the gastrointestinal tract. Boron excretion occurs mainly through the kidney.

Long-term exposure of humans to boron compounds leads to mild gastrointestinal irritation. Inshort- and long-term animal studies and in reproductive studies with rats, testicular atrophy hasbeen observed. Boric acid and borates have not been shown to be mutagenic in various in vitrotest systems. No increase in tumour incidences have been observed in long-term carcinogenicitystudies in mice and rats.

A TDI of 88 g/kg of body weight was derived by applying an uncertainty factor of 100 (for inter-and intraspecies variation) to a NOAEL, for testicular atrophy, of 8.8 mg/kg of body weight perday in a 2-year diet study in dogs. This gives a guideline value for boron of 0.3 mg/litre (roundedfigure), allocating 10% of the TDI to drinking-water. It should be noted, however, that the intake ofboron from food is poorly characterized and that its removal by drinking-water treatment appears

to be poor.

Cadmium

Cadmium metal is used in the steel industry and in plastics. Cadmium compounds are widelyused in batteries. Cadmium is released to the environment in wastewater, and diffuse pollution iscaused by contamination from fertilizers and local air pollution. Contamination in drinking-watermay also be caused by impurities in the zinc of galvanized pipes and solders and some metalfittings, although levels in drinking-water are usually less than 1 g/litre. Food is the main sourceof daily exposure to cadmium. The daily oral intake is 10-35 g. Smoking is a significantadditional source of cadmium exposure.

Absorption of cadmium compounds is dependent on the solubility of the compounds. Cadmium

accumulates primarily in the kidneys and has a long biological half-life in humans of 10-35 years.

There is evidence that cadmium is carcinogenic by the inhalation route, and IARC has classifiedcadmium and cadmium compounds in Group 2A. However, there is no evidence ofcarcinogenicity by the oral route, and no clear evidence for the genotoxicity of cadmium.

The kidney is the main target organ for cadmium toxicity. The critical cadmium concentration inthe renal cortex that would produce a 10% prevalence of low-molecular-weight proteinuria in thegeneral population is about 200 mg/kg, and would be reached after a daily dietary intake of about175 g per person for 50 years.

7/28/2019 2edvol1c

12/65

Assuming an absorption rate for dietary cadmium of 5% and a daily excretion rate of 0.005% ofbody burden, JECFA concluded that, if levels of cadmium in the renal cortex are not to exceed 50mg/kg, the total intake of cadmium should not exceed 1 g/kg of body weight per day. Theprovisional tolerable weekly intake (PTWI) was therefore set at 7 g/kg of body weight. It isrecognized that the margin between the PTWI and the actual weekly intake of cadmium by thegeneral population is small, less than 10-fold, and that this margin may be even smaller insmokers. A guideline value for cadmium of 0.003 mg/litre is established based on an allocation of10% of the PTWI to drinking-water.

Chloride

Chloride in drinking-water originates from natural sources, sewage and industrial effluents, urbanrun-off containing de-icing salt, and saline intrusion.

The main source of human exposure to chloride is the addition of salt to food, and the intake fromthis source is usually greatly in excess of that from drinking-water.

Excessive chloride concentrations increase rates of corrosion of metals in the distribution system,depending on the alkalinity of the water. This can lead to increased concentrations of metals in

the supply.

No health-based guideline value is proposed for chloride in drinking-water. However, chlorideconcentrations in excess of about 250 mg/litre can give rise to detectable taste in water.

Chromium

Chromium is widely distributed in the earth's crust. It can exist in valences of +2 to +6. Totalchromium concentrations in drinking-water are usually less than 2 g/litre, althoughconcentrations as high as 120 g/litre have been reported. In general, food appears to be themajor source of intake.

The absorption of chromium after oral exposure is relatively low and depends on the oxidation

state. Chromium(VI) is more readily absorbed from the gastrointestinal tract than chromium(III)and is able to penetrate cellular membranes.

There are no adequate toxicity studies available to provide a basis for a NOAEL. In a long-termcarcinogenicity study in rats given chromium(III) by the oral route, no increase in tumourincidence was observed. In rats, chromium(VI) is a carcinogen via the inhalation route, althoughthe limited data available do not show evidence for carcinogenicity via the oral route. Inepidemiological studies, an association has been found between exposure to chromium(VI) bythe inhalation route and lung cancer. IARC has classified chromium(VI) in Group 1 (humancarcinogen) and chromium(III) in Group 3.

Chromium(VI) compounds are active in a wide range of in vitro and in vivo genotoxicity tests,whereas chromium(III) compounds are not. The mutagenic activity of chromium(VI) can be

decreased or abolished by reducing agents, such as human gastric juice.

In principle, it was considered that different guideline values for chromium(III) and chromium(VI)should be derived. However, current analytical methods favour a guideline value for totalchromium.

Because of the carcinogenicity of chromium(VI) by the inhalation route and its genotoxicity, thecurrent guideline value of 0.05 mg/litre has been questioned, but the available toxicological datado not support the derivation of a new value. As a practical measure, 0.05 mg/litre, which isconsidered to be unlikely to give rise to significant risks to health, has been retained as the

7/28/2019 2edvol1c

13/65

provisional guideline value until additional information becomes available and chromium can bere-evaluated.

Copper

Copper levels in drinking-water are usually low at only a few micrograms per litre, but copperplumbing may result in greatly increased concentrations. Concentrations can reach severalmilligrams per litre following a period of stagnation in pipes.

Copper is an essential element, and the intake from food is normally 1-3 mg/day. In adults, theabsorption and retention rates of copper depend on the daily intake; as a consequence, copperoverload is unlikely. Acute gastric irritation may be observed in some individuals at concentrationsin drinking-water above 3 mg/litre. In adults with hepatolenticular degeneration, the copperregulatory mechanism is defective, and long-term ingestion can give rise to liver cirrhosis.

Copper metabolism in infants, unlike that in adults, is not well developed, and the liver of thenewborn infant contains over 90% of the body burden, with much higher levels than in adults.Since 1984, there has been some concern regarding the possible involvement of copper fromdrinking-water in early childhood liver cirrhosis in bottle-fed infants, although this has not beenconfirmed.

In 1982, JECFA proposed a provisional maximum tolerable daily intake (PMTDI) of 0.5 mg/kg ofbody weight, based on a rather old study in dogs. With an allocation of 10% of the PMTDI todrinking-water, a provisional health-based guideline value of 2 mg/litre (rounded figure) iscalculated. This study did not take into account the differences in copper metabolism in theneonate. However, a concentration of 2 mg/litre should also contain a sufficient margin of safetyfor bottle-fed infants, because their copper intake from other sources is usually low.

In view of the remaining uncertainties regarding copper toxicity in humans, the guideline value isconsidered provisional. Copper can give rise to taste problems.

Cyanide

The acute toxicity of cyanides is high. Cyanides can be found in some foods, particularly in somedeveloping countries, and they are occasionally found in drinking-water, primarily as aconsequence of industrial contamination.

Effects on the thyroid and particularly the nervous system were observed in some populations asa consequence of the long-term consumption of inadequately processed cassava containing highlevels of cyanide. This problem seems to have decreased significantly in the West Africanpopulations in which it was widely reported, following a change in processing and a generalimprovement in nutritional status.

There are a very limited number of toxicological studies suitable for use in deriving a guidelinevalue. There is, however, some indication in the literature that pigs may be more sensitive thanrats. There is only one study available in which a clear effect level was observed, at 1.2 mg/kg of

body weight per day, in pigs exposed for 6 months. The effects observed were in behaviouralpatterns and serum biochemistry.

Using the LOAEL from this study and applying an uncertainty factor of 100 to reflect inter- andintraspecies variation (no additional factor for a LOAEL was considered necessary because ofdoubts over the biological significance of the observed changes), a TDI of 12 g/kg of bodyweight was calculated.

An allocation of 20% of the TDI to drinking-water was made because exposure to cyanide fromother sources is normally small and because exposure from water is only intermittent. This results

7/28/2019 2edvol1c

14/65

in a guideline value of 0.07 mg/litre (rounded figure), which is considered to be protective foracute and long-term exposure.

Fluoride

Fluorine accounts for about 0.3 g/kg of the earth's crust. Inorganic fluorine compounds are usedin the production of aluminium, and fluoride is released during the manufacture and use ofphosphate fertilizers, which contain up to 4% fluorine.

Levels of daily exposure to fluoride depend on the geographical area. If diets contain fish and tea,exposure via food may be particularly high. In specific areas, other foods and indoor air pollutionmay contribute considerably to total exposure. Additional intake may result from the use offluoride toothpastes.

Exposure to fluoride from drinking-water depends greatly on natural circumstances. Levels in rawwater are normally below 1.5 mg/litre, but ground water may contain about 10 mg/litre in areasrich in fluoride-containing minerals. Fluoride is sometimes added to drinking-water to preventdental caries.

Soluble fluorides are readily absorbed in the gastrointestinal tract after intake in drinking-water.

In 1987, IARC classified inorganic fluorides in Group 3. Although there was equivocal evidence ofcarcinogenicity in one study in male rats, extensive epidemiological studies have shown noevidence of carcinogenicity in humans.

There is no evidence to suggest that the guideline value of 1.5 mg/litre set in 1984 needs to berevised. Concentrations above this value carry an increasing risk of dental fluorosis, and muchhigher concentrations lead to skeletal fluorosis. The value is higher than that recommended forartificial fluoridation of water supplies. In setting national standards for fluoride, it is particularlyimportant to consider climatic conditions, volumes of water intake, and intake of fluoride fromother sources (e.g., food, air). In areas with high natural fluoride levels, it is recognized that theguideline value may be difficult to achieve in some circumstances with the treatment technologyavailable (see section 6.3.5).

Hardness

Hardness in water is caused by dissolved calcium and, to a lesser extent, magnesium. It isusually expressed as the equivalent quantity of calcium carbonate.

Depending on pH and alkalinity, hardness of above about 200 mg/litre can result in scaledeposition, particularly on heating. Soft waters with a hardness of less than about 100 mg/litrehave a low buffering capacity and may be more corrosive to water pipes.

Although a number of ecological and analytical epidemiological studies have shown a statisticallysignificant inverse relationship between hardness of drinking-water and cardiovascular disease,the available data are inadequate to permit a conclusion that the association is causal. There is

some indication that very soft waters may have an adverse effect on mineral balance, but detailedstudies were not available for evaluation.

No health-based guideline value is proposed for hardness. However, the degree of hardness inwater may affect its acceptability to the consumer in terms of taste and scale deposition.

Hydrogen sulfide

Hydrogen sulfide is a gas with an offensive rotten eggs odour that is detectable at very lowconcentrations, below 8 g/m

3in air. It is formed when sulfides are hydrolysed in water. However,

7/28/2019 2edvol1c

15/65

the level of hydrogen sulfide found in drinking-water will usually be low, because sulfides arereadily oxidized in well-aerated water.

The acute toxicity to humans of hydrogen sulfide following inhalation of the gas is high; eyeirritation can be observed at concentrations of 15-30 mg/m

3. Although oral toxicity data are

lacking, it is unlikely that a person could consume a harmful dose of hydrogen sulfide fromdrinking-water. Consequently, no health-based guideline value is proposed. However, hydrogensulfide should not be detectable in drinking-water by taste or odour.

Iron

Iron is one of the most abundant metals in the earth's crust. It is found in natural fresh waters atlevels ranging from 0.5 to 50 mg/litre. Iron may also be present in drinking-water as a result of theuse of iron coagulants or the corrosion of steel and cast iron pipes during water distribution.

Iron is an essential element in human nutrition. Estimates of the minimum daily requirement foriron depend on age, sex, physiological status, and iron bioavailability and range from about 10 to50 mg/day.

As a precaution against storage in the body of excessive iron, in 1983 JECFA established a

provisional maximum tolerable daily intake (PMTDI) of 0.8 mg/kg of body weight, which applies toiron from all sources except for iron oxides used as colouring agents, and iron supplements takenduring pregnancy and lactation or for specific clinical requirements. An allocation of 10% of thisPMTDI to drinking-water gives a value of about 2 mg/litre, which does not present a hazard tohealth. The taste and appearance of drinking-water will usually be affected below this level.

No health-based guideline value for iron in drinking-water is proposed.

Lead

Lead is used principally in the production of lead-acid batteries, solder, and alloys. Theorganolead compounds tetraethyl and tetramethyl lead have also been used extensively asantiknock and lubricating agents in petrol, although their use for these purposes in many

countries is being phased out. Owing to the decreasing use of lead-containing additives in petroland of lead-containing solder in the food processing industry, concentrations in air and food aredeclining, and intake from drinking-water constitutes a greater proportion of total intake.

Lead is present in tapwater to some extent as a result of its dissolution from natural sources, butprimarily from household plumbing systems containing lead in pipes, solder, fittings, or theservice connections to homes. The amount of lead dissolved from the plumbing system dependson several factors, including pH, temperature, water hardness, and standing time of the water,with soft, acidic water being the most plumbosolvent.

Placental transfer of lead occurs in humans as early as the twelfth week of gestation andcontinues throughout development. Young children absorb 4-5 times as much lead as adults, andthe biological half-life may be considerably longer in children than in adults.

Lead is a general toxicant that accumulates in the skeleton. Infants, children up to six years ofage, and pregnant women are most susceptible to its adverse health effects. Inhibition of the

activity ofd-aminolaevulinic dehydratase (porphobilinogen synthase; one of the major enzymesinvolved in the biosynthesis of haem) in children has been observed at blood lead levels as lowas 5 g/dl, although adverse effects are not associated with its inhibition at this level. Lead alsointerferes with calcium metabolism, both directly and by interfering with vitamin D metabolism.These effects have been observed in children at blood lead levels ranging from 12 to 120 g/dl,with no evidence of a threshold.

7/28/2019 2edvol1c

16/65

Lead is toxic to both the central and peripheral nervous systems, inducing subencephalopathicneurological and behavioural effects. There is electrophysiological evidence of effects on thenervous system in children with blood levels well below 30 g/dl. The balance of evidence fromcross-sectional epidemiological studies indicates that there are statistically significantassociations between blood lead levels of 30 g/dl and more and intelligence quotient deficits ofabout four points in children. Results from prospective (longitudinal) epidemiological studiessuggest that prenatal exposure to lead may have early effects on mental development that do notpersist to the age of 4 years. Research on primates has supported the results of theepidemiological studies, in that significant behavioural and cognitive effects have been observedfollowing postnatal exposure resulting in blood lead levels ranging from 11 to 33 g/dl.

Renal tumours have been induced in experimental animals exposed to high concentrations oflead compounds in the diet, and IARC has classified lead and inorganic lead compounds inGroup 2B (possible human carcinogen). However, there is evidence from studies in humans thatadverse neurotoxic effects other than cancer may occur at very low concentrations of lead andthat a guideline value derived on this basis would also be protective for carcinogenic effects.

In 1986, JECFA established a provisional tolerable weekly intake (PTWI) for lead of 25 g/kg ofbody weight (equivalent to 3.5 g/kg of body weight per day) for infants and children on the basisthat lead is a cumulative poison and that there should be no accumulation of body burden of lead.

Assuming a 50% allocation to drinking-water for a 5-kg bottle-fed infant consuming 0.75 litres ofdrinking-water per day, the health-based guideline value is 0.01 mg/litre (rounded figure). Asinfants are considered to be the most sensitive subgroup of the population, this guideline valuewill also be protective for other age groups.

Lead is exceptional in that most lead in drinking-water arises from plumbing in buildings and theremedy consists principally of removing plumbing and fittings containing lead. This requires muchtime and money, and it is recognized that not all water will meet the guideline immediately.Meanwhile, all other practical measures to reduce total exposure to lead, including corrosioncontrol, should be implemented.

Manganese

Manganese is one of the more abundant metals in the earth's crust and usually occurs togetherwith iron. Dissolved manganese concentrations in ground and surface waters that are poor inoxygen can reach several milligrams per litre. On exposure to oxygen, manganese can forminsoluble oxides that may result in undesirable deposits and colour problems in distributionsystems. Daily intake of manganese from food by adults is between 2 and 9 mg.

Manganese is an essential trace element with an estimated daily nutritional requirement of 30-50g/kg of body weight. Its absorption rate can vary considerably according to actual intake,chemical form, and presence of other metals, such as iron and copper, in the diet. Very highabsorption rates of manganese have been observed in infants and young animals.

Evidence of manganese neurotoxicity has been seen in miners following prolonged exposure todusts containing manganese. There is no convincing evidence of toxicity in humans associated

with the consumption of manganese in drinking-water, but only limited studies are available.

Intake of manganese can be as high as 20 mg/day without apparent ill effects. With an intake of12 mg/day, a 60-kg adult would receive 0.2 mg/kg of body weight per day. Allocating 20% of theintake to drinking-water, and applying an uncertainty factor of 3 to allow for possible increasedbioavailability of manganese from water, gives a value of 0.4 mg/litre.

Although no single study is suitable for use in calculating a guideline value, the weight ofevidence from actual daily intake and studies in laboratory animals given manganese in drinking-water in which neurotoxic and other toxic effects were observed supports the view that a

7/28/2019 2edvol1c

17/65

provisional health-based guideline value of 0.5 mg/litre should be adequate to protect publichealth.

It should be noted that manganese may be objectionable to consumers even at levels below theprovisional guideline value.

Mercury

Mercury is present in the inorganic form in surface and ground waters at concentrations usually ofless than 0.5 g/litre. Levels in air are in the range of 2-10 ng/m

3. Mean dietary intake of mercury

in various countries ranges from 2 to 20 g per day per person.

The kidney is the main target organ for inorganic mercury, whereas methyl-mercury affects mainlythe central nervous system.

In 1972, JECFA established a provisional tolerable weekly intake (PTWI) of 5 g/kg of bodyweight of total mercury, of which no more than 3.3 g/kg of body weight should be present asmethylmercury. In 1988, JECFA reassessed methylmercury, as new data had become available,and confirmed the previously recommended PTWI of 3.3 g/kg of body weight for the generalpopulation, but noted that pregnant women and nursing mothers were likely to be at greater risk

from the adverse effects of methylmercury. The available data were considered insufficient toallow a specific methylmercury intake to be recommended for this population group.

To be on the conservative side, the PTWI for methylmercury was used to derive a guideline valuefor inorganic mercury in drinking-water. As the main exposure is from food, a 10% allocation ofthe PTWI to drinking-water was made. The guideline value for total mercury is 0.001 mg/litre(rounded figure).

Molybdenum

Concentrations of molybdenum in drinking-water are usually less than 0.01 mg/litre. However, inareas near mining sites, molybdenum concentrations as high as 200 g/litre have been reported.Dietary intake of molybdenum is about 0.1 mg per day per person. Molybdenum is considered to

be an essential element, with an estimated daily requirement of 0.1-0.3 mg for adults.

No data are available on the carcinogenicity of molybdenum by the oral route. In a 2-year study ofhumans exposed through their drinking-water, the NOAEL was found to be 0.2 mg/litre. There aresome concerns about the quality of this study. An uncertainty factor of 10 would normally beapplied to reflect intraspecies variation. However, as molybdenum is an essential element, afactor of 3 is considered to be adequate. This gives a guideline value of 0.07 mg/litre (roundedfigure).

This value is within the range of that derived on the basis of results of toxicological studies inanimal species and is consistent with the essential daily requirement.

Nickel

The concentration of nickel in drinking-water is normally less than 0.02 mg/litre. Nickel releasedfrom taps and fittings may contribute up to 1 mg/litre. In special cases of release from natural orindustrial nickel deposits in the ground, the nickel concentration in drinking-water may be evenhigher. The average daily dietary intake is normally 0.1-0.3 mg of nickel but may be as high as0.9 mg with an intake of special food items.

The relevant database for deriving a NOAEL is limited. On the basis of a dietary study in rats inwhich altered organ-to-body weight ratios were observed, a NOAEL of 5 mg/kg of body weightper day was chosen. A TDI of 5 g/kg of body weight was derived using an uncertainty factor of

7/28/2019 2edvol1c

18/65

1000: 100 for inter- and intraspecies variation and an extra factor of 10 to compensate for the lackof adequate studies on long-term exposure and reproductive effects, the lack of data oncarcinogenicity by the oral route (although nickel, as both soluble and sparingly solublecompounds, is now considered as a human carcinogen in relation to pulmonary exposure), and amuch higher intestinal absorption when taken on an empty stomach in drinking-water than whentaken together with food.

With an allocation of 10% of the TDI to drinking-water, the health-based guideline value is 0.02mg/litre (rounded figure). This value should provide sufficient protection for individuals who aresensitive to nickel.

Nitrate and nitrite

Nitrate and nitrite are naturally occurring ions that are part of the nitrogen cycle.

Naturally occurring nitrate levels in surface and ground water are generally a few milligrams perlitre. In many ground waters, an increase of nitrate levels has been observed owing to theintensification of farming practice. Concentrations can reach several hundred milligrams per litre.In some countries, up to 10% of the population may be exposed to nitrate levels in drinking-waterof above 50 mg/litre.

In general, vegetables will be the main source of nitrate intake when levels in drinking-water arebelow 10 mg/litre. When nitrate levels in drinking-water exceed 50 mg/litre, drinking-water will bethe major source of total nitrate intake.

Experiments suggest that neither nitrate nor nitrites act directly as a carcinogen in animals, butthere is some concern about increased risk of cancer in humans from the endogenous andexogenous formation of N-nitroso compounds, many of which are carcinogenic in animals.Suggestive evidence relating dietary nitrate exposure to cancer, especially gastric cancer, isavailable from geographical correlation or ecological epidemiological studies, but these resultshave not been confirmed in more definitive analytical studies. It must be recognized that manyfactors in addition to environmental nitrate exposure may be involved.

In summary, the epidemiological evidence for an association between dietary nitrate and canceris insufficient, and the guideline value for nitrate in drinking-water is established solely to preventmethaemoglobinaemia, which depends upon the conversion of nitrate to nitrite. Although bottle-fed infants of less than 3 months of age are most susceptible, occasional cases have beenreported in some adult populations.

Extensive epidemiological data support the current guideline value for nitrate-nitrogen of 10mg/litre. However, this value should not be expressed on the basis of nitrate-nitrogen but on thebasis of nitrate itself, which is the chemical entity of concern to health, and the guideline value fornitrate is therefore 50 mg/litre.

As a result of recent evidence of the presence of nitrite in some water supplies, it was concludedthat a guideline value for nitrite should be proposed. However, the available animal studies are

not appropriate for the establishment of a firm NOAEL for methaemoglobinaemia in rats.Therefore, a pragmatic approach was followed, accepting a relative potency for nitrite and nitratewith respect to methaemoglobin formation of 10:1 (on a molar basis). On this basis, a provisionalguideline value for nitrite of 3 mg/litre is proposed. Because of the possibility of simultaneousoccurrence of nitrite and nitrate in drinking-water, the sum of the ratios of the concentration ofeach to its guideline value should not exceed 1, i.e.

1GV

C

GV

C

nitrate

nitrate

nitrite

nitrite+

7/28/2019 2edvol1c

19/65

where

C= concentrationGV= guideline value.

Dissolved oxygen

No health-based guideline value is recommended for dissolved oxygen in drinking-water.However, a dissolved oxygen content substantially lower than the saturation concentration maybe indicative of poor water quality.

pH

No health-based guideline value is proposed for pH, although eye irritation and exacerbation ofskin disorders have been associated with pH values greater than 11. Although pH usually has nodirect impact on consumers, it is one of the most important operational water quality parameters.

Selenium

Selenium levels in drinking-water vary greatly in different geographical areas but are usuallymuch less than 0.01 mg/litre. Foodstuffs such as cereals, meat, and fish are the principal sourceof selenium in the general population. Levels in food vary greatly according to geographical areaof production.

Selenium is an essential element for humans and forms an integral part of the enzymeglutathione peroxidase and probably other proteins as well. Most selenium compounds are water-soluble and are efficiently absorbed from the intestine. The toxicity of selenium compoundsappears to be of the same order in both humans and laboratory animals.

Except for selenium sulfide, which does not occur in drinking-water, experimental data do notindicate that selenium is carcinogenic. IARC has placed selenium and selenium compounds inGroup 3. Selenium compounds have been shown to be genotoxic in in vitro systems with

metabolic activation, but not in humans. This effect may be dose-dependent in vivo. There is noevidence of teratogenic effects in monkeys, but no data exist for humans.

Long-term toxicity in rats is characterized by depression of growth and liver pathology at seleniumlevels of 0.03 mg/kg of body weight per day given in food.

In humans, the toxic effects of long-term selenium exposure are manifested in nails, hair andliver. Data from China indicate that clinical signs occur at a daily intake above 0.8 mg. Dailyintakes of Venezuelan children with clinical signs were estimated to be about 0.7 mg, on the basisof their blood levels and the Chinese data on the relationship between blood level and intake.Effects on synthesis of a liver protein were also seen in a small group of patients with rheumatoidarthritis given selenium at a rate of 0.25 mg/day in addition to selenium from food. No clinical orbiochemical signs of selenium toxicity were reported in a group of 142 persons with a mean daily

intake of 0.24 mg (maximum 0.72 mg).

On the basis of these data, the NOAEL in humans was estimated to be about 4 g/kg of bodyweight per day. The recommended daily intake of selenium is about 1 g/kg of body weight foradults. An allocation of 10% of the NOAEL in humans to drinking-water gives a health-basedguideline value of 0.01 mg/litre (rounded figure).

Silver

Silver occurs naturally mainly in the form of its very insoluble and immobile oxides, sulfides, and

7/28/2019 2edvol1c

20/65

some salts. It has occasionally been found in ground, surface, and drinking-water atconcentrations above 5 g/litre. Levels in drinking-water treated with silver for disinfection (seesection 6.3.4) may be above 50 g/litre. Recent estimates of daily intake are about 7 g perperson.

Only a small percentage of silver is absorbed. Retention rates in humans and laboratory animalsrange between 0 and 10%.

The only obvious sign of silver overload is argyria, a condition in which skin and hair are heavilydiscoloured by silver in the tissues. An oral NOAEL for argyria in humans for a total lifetime intakeof 10 g of silver was estimated on the basis of human case reports and long-term animalexperiments.

The low levels of silver in drinking-water, generally below 5 g/litre, are not relevant to humanhealth with respect to argyria. On the other hand, special situations exist where silver salts maybe used to maintain the bacteriological quality of drinking-water. Higher levels of silver, up to 0.1mg/litre (this concentration gives a total dose over 70 years of half the human NOAEL of 10 g),could be tolerated in such cases without risk to health.

No health-based guideline value is proposed for silver in drinking-water.

Sodium

Sodium salts (e.g., sodium chloride) are found in virtually all food (the main source of dailyexposure) and drinking-water. Although concentrations of sodium in potable water are typicallyless than 20 mg/litre, they can greatly exceed this in some countries. The levels of sodium salts inair are normally low in relation to those in food or water. It should be noted that some watersofteners can add significantly to the sodium content of drinking-water.

No firm conclusions can be drawn concerning the possible association between sodium indrinking-water and the occurrence of hypertension. Therefore, no health-based guideline value isproposed. However, concentrations in excess of 200 mg/litre may give rise to unacceptable taste.

Sulfate

Sulfates occur naturally in numerous minerals and are used commercially, principally in thechemical industry. They are discharged into water in industrial wastes and through atmosphericdeposition; however, the highest levels usually occur in ground water and are from naturalsources. In general, food is the principal source of exposure to sulfate, although intake fromdrinking-water can exceed that from food in areas with high concentrations. The contribution of airto total intake is negligible.

Sulfate is one of the least toxic anions; however, catharsis, dehydration, and gastrointestinalirritation have been observed at high concentrations. Magnesium sulfate, or Epsom salts, hasbeen used as a cathartic for many years.

No health-based guideline is proposed for sulfate. However, because of the gastrointestinaleffects resulting from ingestion of drinking-water containing high sulfate levels, it is recommendedthat health authorities be notified of sources of drinking-water that contain sulfate concentrationsin excess of 500 mg/litre. The presence of sulfate in drinking-water may also cause noticeabletaste and may contribute to the corrosion of distribution systems.

Inorganic tin

Tin is used principally in the production of coatings used in the food industry. Food, particularlycanned food, therefore represents the major route of human exposure to tin. For the general

7/28/2019 2edvol1c

21/65

population, drinking-water is not a significant source of tin, and levels in drinking-water greaterthan 1-2 g/litre are exceptional. However, there is increasing use of tin in solder, which may beused in domestic plumbing.

Tin and inorganic tin compounds are poorly absorbed from the gastrointestinal tract, do notaccumulate in tissues, and are rapidly excreted, primarily in the faeces.

No increased incidence of tumours was observed in long-term carcinogenicity studies conductedin mice and rats fed stannous chloride. Tin has not been shown to be teratogenic or fetotoxic inmice, rats, and hamsters. In rats, the NOAEL in a long-term feeding study was 20 mg/kg of bodyweight per day.

The main adverse effect on humans of excessive levels of tin in foods (above 150 mg/kg), suchas canned fruit, has been acute gastric irritation. There is no evidence of adverse effects inhumans associated with chronic exposure to tin.

It was concluded that, because of the low toxicity of inorganic tin, a tentative guideline value couldbe derived three orders of magnitude higher than the normal tin concentration in drinking-water.Therefore, the presence of tin in drinking-water does not represent a hazard to human health. Forthis reason, the establishment of a numerical guideline value for inorganic tin is not deemed

necessary.

Total dissolved solids

Total dissolved solids (TDS) comprise inorganic salts (principally calcium, magnesium,potassium, sodium, bicarbonates, chlorides and sulfates) and small amounts of organic matterthat are dissolved in water. TDS in drinking-water originate from natural sources, sewage, urbanrun-off, and industrial wastewater. Salts used for road de-icing in some countries may alsocontribute to the TDS content of drinking-water. Concentrations of TDS in water vary considerablyin different geological regions owing to differences in the solubilities of minerals.

Reliable data on possible health effects associated with the ingestion of TDS in drinking-water arenot available, and no health-based guideline value is proposed. However, the presence of high

levels of TDS in drinking-water may be objectionable to consumers.

Uranium

Uranium is present in the earth's crust, principally in the hexavalent form. It is used primarily as afuel in nuclear energy plants and is introduced into water supplies as a result of leaching fromnatural sources, from mill tailings, from emissions from the nuclear industry, from the combustionof coal and other fuels, and from phosphate fertilizers. Although available information onconcentrations in food and drinking-water is limited, it is likely that food is the principal source ofintake of uranium in most areas.

Uranium accumulates in the kidney, and nephropathy is the primary induced effect in humansand animals. In experimental animals, uranium most commonly causes damage to the proximal

convoluted tubules of the kidney, predominantly in the distal two-thirds. At doses that are not highenough to destroy a critical mass of kidney cells, the effect is reversible, as some of the lost cellsare replaced.

Adequate short- and long-term studies on the chemical toxicity of uranium are not available, andtherefore a guideline value for uranium in drinking-water was not derived. Until such informationbecomes available, it is recommended that the limits for radiological characteristics of uranium beused (see Chapter 4). The equivalent for natural uranium, based on these limits, is approximately140 g/litre.

7/28/2019 2edvol1c

22/65

Zinc

Zinc is an essential trace element found in virtually all food and potable water in the form of saltsor organic complexes. The diet is normally the principal source of zinc. Although levels of zinc insurface and ground water normally do not exceed 0.01 and 0.05 mg/litre, respectively,concentrations in tapwater can be much higher as a result of dissolution of zinc from pipes.

In 1982, JECFA proposed a provisional maximum tolerable daily intake for zinc of 1 mg/kg ofbody weight. The daily requirement for adult men is 15-20 mg/day. It was concluded that, takinginto account recent studies on humans, the derivation of a health-based guideline value is notrequired at this time. However, drinking-water containing zinc at levels above 3 mg/litre may notbe acceptable to consumers.

3.6.2 Organic constituents

Chlorinated alkanes

Carbon tetrachloride

Carbon tetrachloride is used principally in the production of chlorofluorocarbon refrigerants. It is

released into air and water during manufacturing and use. Although available data onconcentrations in food are limited, the intake of carbon tetrachloride from air is expected to bemuch greater than that from food or drinking-water. Concentrations in drinking-water aregenerally less than 5 g/litre.

Carbon tetrachloride has been classified in Group 2B by IARC. It can be metabolized inmicrosomal systems to a trichloromethyl radical that binds to macromolecules, initiating lipidperoxidation and destroying cell membranes. It has been shown to cause hepatic and othertumours in rats, mice, and hamsters after oral, subcutaneous, and inhalation exposure. The timeto first tumour has sometimes been short, within 12-16 weeks in some experiments.

Carbon tetrachloride has not been shown to be mutagenic in bacterial tests with or withoutmetabolic activation, nor has it been shown to induce effects on chromosomes or unscheduled

DNA synthesis in mammalian cells either in vivo or in vitro. It has induced point mutations andgene recombination in a eukaryoric test system.

Carbon tetrachloride, therefore, has not been shown to be genotoxic in most available studies,and it is possible that it acts as a non-genotoxic carcinogen. The NOAEL in a 12-week oralgavage study in rats was 1 mg/kg of body weight per day. A TDI of 0.714 g/kg of body weight(allowing for 5 days per week dosing) was calculated by applying an uncertainty factor of 1000(100 for intra- and interspecies variation, and 10 for evidence of possibly non-genotoxiccarcinogenicity). No additional factor for the short duration of the study was incorporated. It wasconsidered to be unnecessary because the compound was administered in corn oil in the criticalstudy and available data indicate that the toxicity following administration in water may be anorder of magnitude less. The guideline value derived from this TDI, based on 10% allocation todrinking-water, is 2 g/litre (rounded figure).

Dichloromethane

Dichloromethane, or methylene chloride, is widely used as a solvent for many purposes, includingcoffee decaffeination and paint stripping. Exposure from drinking-water is likely to be insignificantcompared with other sources.

Dichloromethane is of low acute toxicity. An inhalation study in mice provided conclusiveevidence of carcinogenicity, whereas a drinking-water study provided only suggestive evidence.IARC has placed dichloromethane in Group 2B; however, the balance of evidence suggests that

7/28/2019 2edvol1c

23/65

it is not a genotoxic carcinogen and that genotoxic metabolites are not formed in relevantamounts in vivo.

A TDI of 6 g/kg of body weight was calculated by applying an uncertainty factor of 1000 (100 forinter- and intraspecies variation and 10 reflecting concern for carcinogenic potential) to a NOAELof 6 mg/kg of body weight per day for hepatotoxic effects in a 2-year drinking-water study in rats.This gives a guideline value of 20 g/litre (rounded figure), allocating 10% of the TDI to drinking-water. It should be noted that widespread exposure from other sources is possible.

1,1-Dichloroethane

1,1-Dichloroethane is used as a chemical intermediate and solvent. There are limited datashowing that it can be present in concentrations of up to 10 g/litre in drinking-water. However,because of the widespread use and disposal of this chemical, its occurrence in ground water mayincrease.

1,1-Dichloroethane is rapidly metabolized by mammals to acetic acid and a variety of chlorinatedcompounds. It is of relatively low acute toxicity, and limited data are available on its toxicity fromshort- and long-term studies.

There is limited in vitro evidence of genotoxicity. One carcinogenicity study by gavage in miceand rats provided no conclusive evidence of carcinogenicity, although there was some evidenceof an increased incidence of haemangiosarcomas in treated animals.

In view of the very limited database on toxicity and carcinogenicity, it was concluded that noguideline value should be proposed.

1,2-Dichloroethane

1,2-Dichloroethane is used mainly as an intermediate in the production of vinyl chloride and otherchemicals and to a lesser extent as a solvent. It has been found in drinking-water at levels of upto a few micrograms per litre. It is found in urban air.

IARC has classified 1,2-dichloroethane in Group 2B. It has been shown to produce statisticallysignificant increases in a number of tumour types in laboratory animals, including the relativelyrare haemangiosarcoma, and the balance of evidence indicates that it is potentially genotoxic.There are no suitable long-term studies on which to base a TDI.

On the basis of haemangiosarcomas observed in male rats in a 78-week gavage study, andapplying the linearized multistage model, a guideline value for drinking-water of 30 g/litre,corresponding to an excess lifetime cancer risk of 10

-5, was calculated.

1,1,1-Trichloroethane

1,1,1-Trichloroethane has been found in only a small proportion of surface and ground waters,usually at concentrations of less than 20 g/litre. In a few instances, much higher concentrations

have been observed. There appears to be increasing exposure to 1,1,1-trichloroethane.

1,1,1-Trichloroethane is rapidly absorbed from the lungs and gastrointestinal tract, but only smallamounts - about 6% in humans and 3% in experimental animals - are metabolized. Exposure tohigh concentrations can lead to hepatic steatosis (fatty liver) in both humans and laboratoryanimals.

IARC has placed 1,1,1-trichloroethane in Group 3. Available studies of oral administration wereconsidered inadequate for calculation of a TDI. As there is an increasing need for guidance onthis compound, a 14-week inhalation study in male mice was selected for use in calculating the

7/28/2019 2edvol1c

24/65

guideline value. Based on a NOAEL of 1365 mg/m3, a TDI of 580 g/kg of body weight was

calculated from a total absorbed dose of 580 mg/kg of body weight per day (assuming anaverage mouse body weight of 30 g, breathing rate of 0.043 m

3/day, and absorption of 30% of the

air concentration), applying an uncertainty factor of 1000 (100 for inter- and intraspecies variationand 10 for the short duration of the study). A provisional guideline value of 2000 g/litre (roundedvalue) is proposed, allocating 10% of the TDI to drinking-water.

This value is provisional because of the use of an inhalation study rather than an oral study. It isstrongly recommended that an adequate oral toxicity study be conducted to provide moreacceptable data for the derivation of a guideline value.

Chlorinated ethenes

Vinyl chloride is used primarily for the production of polyvinyl chloride. The background level ofvinyl chloride in ambient air in western Europe is estimated to range from 0.1 to 0.5 g/m

3.

Residual vinyl chloride levels in food and drinks are now below 10 g/kg. Vinyl chloride has beenfound in drinking-water at levels of up to a few micrograms per litre, and, on occasion, muchhigher concentrations have been found in ground water. It can be formed in water fromtrichloroethene and tetrachloroethene.

Vinyl chloride is metabolized to highly reactive and mutagenic metabolites by a dose-dependentand saturable pathway.

The acute toxicity of vinyl chloride is low, but vinyl chloride is toxic to the liver after short- andlong-term exposure to low concentrations. Vinyl chloride has been shown to be mutagenic invarious test systems in vitro and in vivo.

There is sufficient evidence of the carcinogenicity of vinyl chloride in humans from industrialpopulations exposed to high concentrations, and IARC has classified vinyl chloride in Group 1. Acausal association between vinyl chloride exposure and angiosarcoma of the liver is sufficientlyproved. Some studies suggest that vinyl chloride is also associated with hepatocellularcarcinoma, brain tumours, lung tumours, and malignancies of the lymphatic and haematopoietictissues.

Animal data show vinyl chloride to be a multisite carcinogen. Vinyl chloride administered orally orby inhalation to mice, rats, and hamsters produced tumours in the mammary gland, lungs,Zymbal gland, and skin, as well as angiosarcomas of the liver and other sites.

Because there are no data on carcinogenic risk following oral exposure of humans to vinylchloride, estimation of risk of cancer in humans was based on animal carcinogenicity bioassaysinvolving oral exposure. Using results from the rat bioassay, which yields the most protectivevalue, and applying the linearized multistage model, the human lifetime exposure for a 10

-5

excess risk of hepatic angiosarcoma was calculated to be 20 g per person per day. It was alsoassumed that, in humans, the number of cancers at other sites may equal that of angiosarcomaof the liver, so that a correction (factor of 2) for cancers other than angiosarcoma is justified.Using the lifetime exposure of 20 g per person per day for a 10

-5excess risk of hepatic

angiosarcoma, a guideline value for drinking-water of 5 g/litre was calculated.

1,1-Dichloroethene

1,1-Dichloroethene, or vinylidene chloride, is an occasional contaminant of drinking-water. It isusually found together with other chlorinated hydrocarbons. There are no data on levels in food,but levels in air are generally less than 40 ng/m

3except at some manufacturing sites.

Following oral or inhalation exposure, 1,1-dichloroethene is almost completely absorbed,extensively metabolized, and rapidly excreted. It is a central nervous system depressant and may

7/28/2019 2edvol1c

25/65

cause liver and kidney toxicity in occupationally exposed humans. It causes liver and kidneydamage in laboratory animals.

IARC has placed 1,1-dichloroethene in Group 3. It was found to be genotoxic in a number of testsystems in vitro but was not active in the dominant lethal assay in vivo. It induced kidney tumoursin mice in one inhalation study but was reported not to be carcinogenic in a number of otherstudies, including several in which it was given in drinking-water.

A TDI of 9 g/kg of body weight was calculated from a LOAEL of 9 mg/kg of body weight per dayin a 2-year drinking-water study in rats, using an uncertainty factor of 1000 (100 for intra- andinterspecies variation and 10 for the use of a LOAEL in place of a NOAEL and the potential forcarcinogenicity). This gives a guideline value of 30 g/litre (rounded figure) for a 10% contributionto the TDI from drinking-water.

1,2-Dichlorethene

1,2-Dichlorethene exists in a cis and a trans form. The cis form is more frequently found as awater contaminant. The presence of these two isomers, which are metabolites of otherunsaturated halogenated hydrocarbons in wastewater and anaerobic ground water, may indicatethe simultaneous presence of more toxic organochlorine chemicals, such as vinyl chloride.

Accordingly, their presence indicates that more intensive monitoring should be conducted. Thereare no data on exposure from food. Concentrations in air are low, with higher concentrations, inthe microgram per cubic metre range, near production sites. The cis-isomer was previously usedas an anaesthetic.

There is little information on the absorption, distribution, and excretion of 1,2-dichloroethene.However, by analogy with 1,1-dichloroethene, it would be expected to be readily absorbed,distributed mainly to the liver, kidneys, and lungs, and rapidly excreted. The cis-isomer is morerapidly metabolized than the trans-isomer in in vitro systems.

Both isomers have been reported to cause increased serum alkaline phosphatase levels inrodents. In a 3-month study in mice given the trans-isomer in drinking-water, there was a reportedincrease in serum alkaline phosphatase and reduced thymus and lung weights. Transient

immunological effects were also reported, the toxicological significance of which is unclear.Trans-1,2-dichloroethene also caused reduced kidney weights in rats, but at higher doses. Onlyone rat toxicity study is available for the cis-isomer, which produced toxic effects in rats similar inmagnitude to those induced by the trans-isomer in mice, but at higher doses.

There are limited data to suggest that both isomers may possess some genotoxic activity. Thereis no information on carcinogenicity.

Data on the trans-isomer were used to calculate a joint guideline value